In this study we examined the question of how error correction occurs in an ensemble of deep convolutional networks, trained for an important applied problem: segmentation of Electrocardiograms(ECG). We also explore the possibility of using the information about ensemble errors to evaluate a quality of data representation, built by the network. This possibility arises from the effect of distillation of outliers, which was demonstarted for the ensemble, described in this paper.
Using electrocardiograms as an example, we demonstrate the characteristic problems that arise when modeling one-dimensional signals containing inaccurate repeating pattern by means of standard convolutional networks. We show that these problems are systemic in nature. They are due to how convolutional networks work with composite objects, parts of which are not fixed rigidly, but have significant mobility. We also demonstrate some counterintuitive effects related to generalization in deep networks.
The Orenburg oil region is one of the oldest regions of the Volga-Ural oil and gas province. The story of the field discoveries in the Orenburg region starts from the end of 30s of the twentieth century. By now the explored stock of large and easy-to-recover deposits is nearly depleted. The production has gradually led to the depletion of the existing deposits, and the replacement of oil producer resource base has become necessary.Since 2014 up to date Tyumen Petroleum Research Center has been carrying out the scientific research for the Orenburgneft producing fields; this research is focused on the detailed studies of oil and gas bearing capacity of the Palaeozoic sedimentary cover and the preparation of geological basis for the determination of prospective areas for exploration.The article presents the results of conducted work on the prospectivity assessment for fifty Orenburgneft fields located in the central and southern parts of the Orenburg region ( Figure 1). The study area includes Bobrovsko-Pokrovsky (BPS), Mikhailovo-Kokhanovsky (MKS) and Bolshekinelsky (BS) swells, Kamelik-Chaganskaya Dislocation system (KCDS), Pavlovskaya saddle (PS) and Vostochno-Orenburgskoye arched uplift (VOAU).Regardless the extensive petroleum history in the region, the potential of new discoveries has not been exhausted yet. Often new discoveries are made accidentally when drilling or well logging. Established concepts and conventional approach to the estimation of oil and gas bearing capacity of the fields lead to under-estimation of productive potential of the prospects.It has become urgent to create common geological database and reconsider all accumulated information that will favor targeted searching of new deposits overlooked by various reasons.There can be several reasons for underestimation of reservoir potential. Among them are insufficient geological studies of the prospects, low informativity of well logs in old wells, insufficient seismic coverage. Another important factor was the economical criterion of production rate profitability in the period of well drilling.
We describe how some problems (interpretability, lack of object-orientedness) of modern deep networks potentially could be solved by adapting a biologically plausible saccadic mechanism of perception. A sketch of such a saccadic vision model is proposed. Proof of concept experimental results are provided to support the proposed approach.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.